Don't Treat Your Citizens Like Criminals!

Jonathan Care

Description

Short Summary

Interesting Facts

Notable Quotes

Recommendations

Takeaways

The UK has granted DWP and HMRC powers to surveil and seize funds from citizens' bank accounts without judicial oversight or reasonable suspicion. Four comparable systems have failed catastrophically: Post Office Horizon (736 wrongful convictions, 4 deaths), HMRC Loan Charge (10 suicides), Australian Robodebt (unlawful, AUD $1.8B cost), Dutch toeslagenaffaire (cabinet resigned, "institutional racism"). All shared features now present in UK proposals: algorithmic opacity, reversed burden of proof, no independent oversight. DWP's own assessments show algorithmic bias by age, disability, and nationality. Without reform: preventable deaths, mass wrongful investigations, discrimination claims, and compensation costs exceeding any fraud recovery. How can fraud be addressed without repeating the largest miscarriage of justice in British legal history? To make this system safe for use, there should be a restoration of judicial authorisation, reasonable suspicion as a threshold, the algorithm(s) should be publicly available for independent audit, and oversight committees with power to halt the system if harm thresholds are exceeded.

Two UK legislative changes are described as quietly inverting a foundational liberal-democratic default: the presumption of innocence in the citizen–state relationship. The Public Authorities (Fraud, Error and Recovery) Act (Royal Assent December 2025) enables the Department for Work and Pensions (DWP) to run the “Eligibility Verification Measure,” requiring fifteen major banks to check the accounts of people receiving specified benefits against undisclosed “eligibility indicators,” without any need for suspicion or “reasonable grounds.” In parallel, HMRC’s Direct Recovery of Debts powers (relaunched September 2025) allow HMRC to freeze and transfer funds directly from individuals’ bank accounts to settle alleged tax debts above £1,000 without court authorisation, relying mainly on internal review. The text argues these shifts create a two-tier model of citizenship in which benefit recipients and individual taxpayers are exposed to intrusive, automated enforcement, while powerful actors (e.g., large companies) face negotiated settlement. Beyond civil liberties, the central concern is systemic risk: opaque, biased, high-impact algorithmic decisioning and a centralised financial “honeypot” with unprecedented attack surface. Multiple past failures are presented as warnings—Post Office Horizon (wrongful convictions and suicides under assumed-reliable computer evidence and supplier secrecy), the Loan Charge (suicides following aggressive retrospective tax demands), Australia’s Robodebt (unlawful automated debts and suicides, massive remediation costs), and the Netherlands’ childcare benefits scandal (algorithmic profiling, mass harm, cabinet resignation). The text concludes the projected financial upside is small relative to total fraud/error losses, while the risk of error, bias, insecurity, and irreversible harm is high, and it recommends restoring due-process safeguards (judicial authorisation, reasonable suspicion), separating fraud from error, mandating transparency and independent audit (including supplier disclosure), creating meaningful pre-action appeals, and imposing independent security review plus an operational “circuit breaker” to halt the system when harms exceed a defined threshold.